最佳PCA稀疏表示方法及在直升机旋翼故障识别中的应用

A global sparse representation scheme based on PCA for fault identification of helicopter rotor.Computer Engineering and Applications

提出一种基于PCA稀疏表示算法进行直升机旋翼故障识别的方法。首先相应的PCA预处理本身计算复杂度不高且能对样本的降维程度较高,其次根据样本相似性原则,基于PCA的稀疏表示方法不仅能保持样本在处理前后相互距离不变,而且提高了计算效率。采用新的诊断模型对直升机旋翼故障分类识别,并与基于神经网络和基于支持向量机的诊断方法进行比较。结果表明本文方法对旋翼故障具有良好的识别能力。

This paper describes an application of global sparse representation scheme based on PCA to fault identiifcation of helicopter rotor.The corresponding PCA ifrst have not high computational complexity and higher sample dimension reduction degree.According to the similarity principle,the sparse representation scheme based on PCA can only keep samples in mutual distance unchanged before and after processing,and improve the computational efifciency.The improved diagnosis model is employed to identify the rotor faults and compared with the diagnosis methods based on neural network and SVM.The results show that the proposed scheme is effective in diagnosing the faults of helicopter rotor.